Log into your member account to listen to this article. Not a member? Join the herd.

The information disorder (i.e., mis- and disinformation) pervasive on social media has arguably interfered with democratic processes across the world. As public authorities and political actors continue to embrace social media as a broadcast and civic engagement tool, the potency of manipulated narratives online is further entrenched. This is debatably truer in electoral contexts where issues are perhaps more emotive and divisive. For example, in the run-up to Kenya’s general elections, a notable amount of mis- and disinformation on social media was observable. As Wambui Wamunyu and June Okal noted, doctored images of crowds during political rallies, mild deepfake videos, doctored photos, and fake accounts passing off as political actors or mainstream media were just some of the categories of mis- and disinformation observable on social media. These observations tie in with earlier research by Odanga Madung and Brian Obilo, highlighting the practice of using bloggers for disinformation campaigns. During the actual elections, the EU Election Observer Mission also observed “manufactured amplification and coordination of messages online by fake accounts and malicious, bot-driven activity in support of the presidential candidates”.

The impact of the information disorder on democracies has been extensively discussed and will not be the subject of this article. Instead, this article focuses on the diverse responses which have been mooted and implemented in Kenya by policy makers, media, civil society, and social media platforms in response to the information disorder. In particular, this article argues that these responses are largely ineffective when used in isolation and suggests that collectives comprised of a broad range of multidisciplinary stakeholders working collaboratively are likely to have a higher chance at success. One such collective, Fumbua, was established in the run-up to the 2022 general elections in Kenya, and this article argues that the frameworks for collaboration it established can be repurposed to address the information disorder in numerous contexts.

Contextualizing the information disorder on social media

The proliferation of mis- and disinformation on social media is made easier by the fact that such platforms, by nature, enable peer-to-peer engagement with little to no gatekeeping. While this characteristic has also meant that these platforms have served to create room for civic engagement and act as an equalizer, such civic engagement is often undermined by the harmful content that is prevalent. In recognition of the potential for harm their platforms pose to democratic processes, numerous social media platforms have adopted policies and tools specifically designed to address election-related mis-and disinformation. Comparatively, the content moderation tools applied in the Global South have arguably been scant. For example, in Brazil, the individuals tasked with enforcing Twitter’s policies during the presidential election only got access to the necessary internal tools a day prior to the election, and only in a limited capacity. Twitter allegedly utilized automatic enforcement technology and third-party service providers. According to numerous commentators, it is not uncommon for content moderation efforts in the Global South to be below par. From automatic enforcement tools trained on datasets lacking in local context, to human content moderators facing the same challenge, these platforms’ efforts to curb the information disorder are handicapped from the outset. These challenges are exacerbated in electoral contexts. Recent developments have shown that it sometimes takes third parties such as researchers or civil society pointing out harmful content for platforms to act.

It is generally agreed that mis-and disinformation was prevalent on social media during Kenya’s August 2022 general election. For example, the EU Election Observer Mission indicated in its report that it had identified hundreds of misleading Facebook and Twitter profiles. Platforms triggered their civic integrity policies a few weeks prior to the election and set up information centres and moderately labelled misleading content. However, these labels were not consistently applied and were in fact only deployed during the election tallying process. Stakeholders seemingly lacked a clear solution to address the information disorder on social media. The lack of sustainable and scalable solutions is not unique to Kenya and the region. It is certainly a global problem and a key step in the right direction is securing more transparency from platforms in relation to their enforcement processes as this will enable stakeholders to co-create solutions. However, in the interim, the information disorder can be addressed by effecting incremental and sustainable changes to how media is produced and consumed. One way to accomplish this is through multidisciplinary collectives such as Fumbua.

Addressing the information disorder

Fumbua is a collective of media and media-related organisations which came together in the run-up of the 2022 general election with a view to addressing the information disorder as it relates to political campaigning. The efforts to address the information disorder in Kenya’s 2022 general election can largely be categorized into actions taken in anticipation of the mis- and disinformation (pre-emptive measures) and actions taken in response to the information disorder (reactive measures). Fumbua brought together organisations involved in both areas, such as fact checkers, “pre-bunkers” and traditional media. Both these reactive and pre-emptive measures are discussed below.

The information disorder can be addressed by effecting incremental and sustainable changes to how media is produced and consumed.

Mis- and disinformation has reportedly featured in Kenyan elections since 2013. Consequently, with each passing cycle, stakeholders have been able to understand its nature and develop solutions which are alive to Kenya’s specific context. Unfortunately, due to the rapidly evolving nature of mis- and disinformation practices, the solutions developed have often been reactive in that they seek to get rid of such harmful content or undo its effects after the fact. For example, by criminalizing false content through the Computer Misuse and Cybercrimes Act, by fact checking such content, by using labels to warn audiences of the nature of the content, and by obtaining the takedown of such content from social media sites.

Fact checking has perhaps been the most prevalent or visible response to the information disorder. It essentially entails systematically breaking down the validity of claims made by public officials, institutions, and political actors with a view to identifying whether the claim is factual or not. In Kenya’s elections, various fact-checkers were active. These included independent media, the fact-checking desks of mainstream media, and collectives or associations. To name a few, Africa Check, Africa Uncensored, Pesa Check, Media Council of Kenya, Kenya Editors Guild and The Star were involved in fact-checking claims made during the Kenyan elections. While fact-checking has increasingly become common, it would be improper to conflate its growing prominence with its ability to address the information disorder, especially when empirical evidence on the subject is divided. In highly politicized environments, it is unlikely that being exposed to verification of claims will affect an audience’s world view. This is more so the case where the objectivity and impartiality of the fact-checkers are in question. Fact-checkers often must compete with an audience’s confirmation bias and their credibility is often questioned due to the conflict their narrative poses to the world view of some audiences. This is not made easier by the fact that fact-checking is a difficult, time-consuming, and labour-intensive process which cannot compete with the speed at which false information is spread through social media. Add in the fact that false information more easily captures attention due to its ability to trigger negative emotions and one can understand why the utility and efficacy of fact-checking is limited. Fact-checking claims made through social media have also been especially difficult in Kenya due to the minimal and often performative support given to fact-checkers by social media platforms.

For fact-checking to be effective, it must offer an alternative narrative to that which it is disputing. The challenge is that such a narrative must exist in the first place and must be capable of being accepted by an audience. Where such a narrative exists, there is a risk that it may come with “political baggage” and as such be difficult to accept. In such cases, the efficacy of fact-checking is limited, and this is essentially the challenge faced by fact-checkers – purveyors of false information are not bound by the same rules. Despite all this, fact-checking has been found to positively affect audience beliefs notwithstanding pre-existing beliefs and whether an alternative narrative was presented. However, these credentials are limited as the effects on belief are weak and gradually becoming negligible. Additionally, they do not always translate to downstream effects (i.e., changing of votes).

For a long time, stakeholders seeking to curb the information disorder have found themselves on the back-foot, always responding after the fact. By the time interventions such as fact-checks, social media takedowns, and flags are deployed, harmful content has likely taken root. With this in mind, some pre-emptive solutions have been contemplated and used by stakeholders. These are discussed below.

While fact-checking has increasingly become common, it would be improper to conflate its growing prominence with its ability to address the information disorder.

As discussed earlier in this article, fact-checkers often face the challenge of having to overcome an audience’s inherent biases and the political baggage accompanying the alternative narratives they seek to put forth. In seeking to overcome this reactionary approach, Stephan Lewandowsky and Sander van der Linden argue that it may be more effective to inoculate audiences against harmful content by priming their minds to anticipate it. This has come to be referred to as prebunking, and it essentially entails exposing audiences to watered down versions of false or misleading content with a view to highlighting the tactics used by purveyors of such content. Prebunking efforts recognize that the information disorder may not necessarily be solved by disseminating more accurate information given that harmful content is often consumed in highly politicized contexts. Instead, these efforts seek to redesign information architecture through behavioural interventions (i.e., changing how audiences consume information). In Kenya, Stop Reflect Verify was the first publicly documented election-related prebunking program. It offered a misinformation quiz focused on the Kenyan elections.

While prebunking seemingly promises to reduce the reactionary nature of stakeholder efforts, there is insufficient proof that skills learned in prebunking programmes are applied in practical situations. Counterfactual thinking may be a useful strategy to incorporate into prebunking efforts. Counterfactual thinking involves stimulating an audience’s mind to consider alternative facts and hypotheses when presented with information in a bid to logically deduce the likely truth. The lack of consensus on the utility and efficacy of prebunking as an alternative to fact-checking points to the need for the deployment of multiple interventions in a coordinated fashion, and this is where multidisciplinary collectives such as Fumbua come in.

Building in sustainability 

Periodically, civil society, media practitioners, and the donor community focus their efforts on election-related programmes in a collaborative manner (for example the media’s collaboration during presidential debates). In most cases, the collaboration does not survive the post-election period. As a result, these election stakeholders have to start anew during each election. A considerable amount of time and resources are dedicated to establishing the frameworks for collaboration, taking away from the potential impact these programmes may have. With collectives such as Fumbua, stakeholders are able to repurpose the goodwill that fostered collaboration during elections to continue to address the information disorder in other contexts. By sustaining the collaboration, stakeholders would be able to leverage on incremental gains and make a more impactful change. In relation to the information disorder, they would be better able to move towards how media is generated and consumed. The effect of this would be a more informed audience that is less susceptible to mis- and disinformation.